Lignocellulose molecular assembly and deconstruction properties of lignin-altered rice mutants

Abstract

Bioengineering approaches to modify lignin content and structure in plant cell walls have shown promise for facilitating biochemical conversions of lignocellulosic biomass into valuable chemicals. Despite numerous research efforts, however, the effect of altered lignin chemistry on the supramolecular assembly of lignocellulose and consequently its deconstruction in lignin-modified transgenic and mutant plants is not fully understood. In this study, we aimed to close this gap by analyzing lignin-modified rice (Oryza sativa L.) mutants deficient in 5-HYDROXYCONIFERALDEHYDE O-METHYLTRANSFERASE (CAldOMT) and CINNAMYL ALCOHOL DEHYDROGENASE (CAD). A set of rice mutants harboring knockout mutations in either or both OsCAldOMT1 and OsCAD2 was generated in part by genome editing and subjected to comparative cell wall chemical and supramolecular structure analyses. In line with the proposed functions of CAldOMT and CAD in grass lignin biosynthesis, OsCAldOMT1-deficient mutant lines produced altered lignins depleted of syringyl and tricin units and incorporating noncanonical 5-hydroxyguaiacyl units, whereas OsCAD2-deficient mutant lines produced lignins incorporating noncanonical hydroxycinnamaldehyde-derived units. All tested OsCAldOMT1- and OsCAD2-deficient mutants, especially OsCAldOMT1-deficient lines, displayed enhanced cell wall saccharification efficiency. Solid-state nuclear magnetic resonance (NMR) and X-ray diffraction analyses of rice cell walls revealed that both OsCAldOMT1- and OsCAD2 deficiencies contributed to the disruptions of the cellulose crystalline network. Further, OsCAldOMT1 deficiency contributed to the increase of the cellulose molecular mobility more prominently than OsCAD2 deficiency, resulting in apparently more loosened lignocellulose molecular assembly. Such alterations in cell wall chemical and supramolecular structures may in part account for the variations of saccharification performance of the OsCAldOMT1- and OsCAD2-deficient rice mutants.

Figure 1

Generation and phenotypic characterization of OsCAldOMT1- and OsCAD2-knockout rice mutant lines. A, gene structures and mutations in the mutant lines. CRISPR/Cas9-induced mutation sites in OsCAldOMT1-knockout lines (caldomt1, cad2 caldomt1-a, and cad2 caldomt1-b) and Tos17-retrotransposon insertion sites in OsCAD2-knockout lines (cad2, cad2 caldomt1-a, and cad2 caldomt1-b) are shown. PAM sites (navy) and deletion or insertion mutation sites (blue) are indicated. B and C, growth characteristics (B) and morphological phenotype (C) of wild-type and homozygous mutant lines at the ripening stage. Bars in (C) = 10 cm. Values refer to mean ± standard deviation (sd) from individually analyzed plants (n = 5). Different letters indicate significant differences (ANOVA with post-hoc Tukey–Kramer’s test, P < 0.05). WT, wild-type control line; caldomt1, OsCAldOMT1 single-knockout line; cad2, OsCAD2 single-knockout line; cad2 caldomt1-a and cad2 caldomt1-b; OsCAldOMT1 and OsCAD2 double-knockout lines.

Figure 2

Lignin content and composition analyses of OsCAldOMT1 and OsCAD2-knockout rice mutant lines based on chemical analyses. A, lignin content determined by a Klason lignin assay. B–F, thioacidolysis-derived lignin composition data. Total yield of p-hydroxyphenyl (H)-, guaiacyl (G)-, and syringyl (S)-type monomers (B), H/G/S monomer composition (C), S/G monomer ratio (D), normalized GC/MS peak area of 5H-type monomers (E) and structures of H-, G-, S-, and 5H-type monomers (trimethylsilylated) released upon thioacidolysis (F) are shown. Values refer to mean ±  sd from individually analyzed plants (n = 3). Different letters indicate significant differences (ANOVA with post-hoc Tukey–Kramer’s test, P < 0.05).

Figure 3

Lignin composition analysis of OsCAldOMT1- and OsCAD2-knockout rice mutant lines based on solution-state 2D NMR. A, aromatic sub-regions of HSQC NMR spectra of dioxane-soluble lignin samples from OsCAldOMT1- and OsCAD2-knockout rice mutant lines. Contour coloration matches the substructures shown. B, normalized signal intensity values and ratio of major lignin aromatic units. Data are expressed on a G₂ + ½S_2/6 + Gʹ₂ + ½Sʹ_2/6 = 1 basis. The dioxane-soluble lignin samples were prepared from culm cell wall residues pooled from three independent plants for each mutant and wild-type line. n.d., not detected.

Figure 4

Lignin inter-monomeric and end-unit linkage analysis of OsCAldOMT1- and OsCAD2-knockout rice mutant lines based on solution-state 2D NMR. A and B, oxygenated-aliphatic (A) and aldehyde (B) sub-regions of HSQC NMR spectra of dioxane-soluble lignin samples from OsCAldOMT1- and OsCAD2-knockout rice mutant lines. Contour coloration matches the substructures shown in each panel. C, normalized signal intensity values of major lignin inter-monomeric linkage types. Data are expressed on I_α + II_α + ½III_α + Iʹ_α + Iʹʹ₉ = 1 basis. The dioxane-soluble lignin samples were prepared from culm cell wall residues pooled from three independent plants for each mutant and wild-type line.

Figure 5

Enzymatic saccharification efficiency of culm cell walls from OsCAldOMT1- and OsCAD2-knockout rice mutant lines. Data are expressed as glucose yield per cell wall residue, CWR (upper) and glucose yield per total glucan (lower). Values refer to mean ±  sd from individually analyzed plants (n = 3). Different letters indicate significant differences (ANOVA with post-hoc Tukey–Kramer’s test, P < 0.05).

Figure 6

Polysaccharide assembly and mobility analyses of culm cell walls from OsCAldOMT1- and OsCAD2-knockout rice mutant lines. A, expanded ¹³C CP–MAS NMR spectra showing crystalline/internal (C_4a) and amorphous/surface (C_4b) cellulose C4 signals. Inserted values are volume integrals for indicated peak areas (C_4a + C_4b = 100). B, apparent cellulose crystallinity index based on WAXD. Values refer to mean ±  sd from individually analyzed plants (n = 3). Different letters indicate significant differences (ANOVA with post-hoc Tukey–Kramer’s test, P < 0.05). C, CP ¹³C spin-lattice relaxation time (T₁) data for major cellulose carbon sites. Delay time-dependent signal decay data were fitted using a double exponential function to determine two independent T₁ for slower and faster relaxing components (see Supplemental Methods). T₁ values (left) and fraction (slower + faster-relaxing components = 100%) data (right) for slower-relaxing components are shown. All data including T₁ and fraction data for faster-relaxing components are listed in Supplemental Table S6. The error bars indicate sds of the fitting coefficients.

Figure 7

Simon’s staining assay of culm cell walls from OsCAldOMT1- and OsCAD2-knockout rice mutant lines. Amounts of absorbed DO and DB dyes (A) and ratio of absorbed DO to DB dyes (B) are shown. Values are means ±  sd from individually analyzed plants (n = 3). Different letters indicate significant differences (ANOVA with post-hoc Tukey–Kramer’s test, P < 0.05).